Bag of Visual Words Model with Deep Spatial Features for Geographical Scene Classification

Table 3

The classification accuracy of different methods on 15-scene and MIT indoor datasets.

Number	Method	MIT indoor (%)	15-scene (%)
1	SPM [4]	34.40	81.4
2	OTC [30]	47.33	—
3	LPR [31]	44.8	85.8
4	Discriminative patches ++ [32]	49.40	—
5	FV + bag of parts [33]	63.18	—
6	MOP-CNN [25]	68.88	—
7	Our approach	66.09	90.1